Gregory Maxwell (gmaxwell(a)gmail.com) [050523 09:55]:
On 5/22/05, David Gerard
<fun(a)thingy.apana.org.au> wrote:
> No, no, only for 1.5 - so that the data gathering
can proceed without any
> visible output as a running score.
> I anticipate the results of the data-gathering test to be available in a
> downloadable form for people to pick over in close detail.
Eh, it sets a bad precedent. Users will claim we stole
their anonymity
when we turn off the feature. Besides what will it hurt? sure it might
get used in fodder for silly arguments but so does everything else.
To make the test the most valid we should try to make it approximate
the real usage as close as possible. For example, it's likely that
people will vote more often when they disagree with the existing
votes. But that wont happen if the existing votes are private.
The thing is we don't know what the real usage will be until we have some
data. An "About the ratings trial" link may be applicable. Then they can't
claim they weren't warned, just because they didn't see the bit on the
twelfth page in <font face="Flyspeck" size="-3">.
I dunno. I was worried about observer effect - particularly with the live
updated ratings tallies - but then, that may also be a test of just how
spammy their little hearts are. Also, that way we don't have to change the
current code and can bug Magnus about other aspects of it instead ;-)
People will claim many things. We are going into 1.5
with an explicit
statement that it's a test and shouldn't be used for anything.
True. People out to do something obnoxious will find an excuse.
> It provides some framework for article metadata,
yes. Or at least for ideas
> on metadata.
Well version metadata. We already have article
metadata of a sort,
think catagory tags and the like. :)
Mmm. Those, like interwiki links, are pseudo-metadata - they work like
metadata, but they're actually in the wikitext, which is not ideal in many
ways.
- d.